Fermi Inc. (FRMI) Earnings Call Transcript

Good morning, everyone, and thank you for joining us for Fermi America's Third Quarter 2026 Earnings Call. My name is Rodrigo Acuna, and I serve as Director of Investor Relations. Joining me on the call today are Lee McIntire, our Chief Executive Officer, Rob Masson, our Chief Financial Officer, and Anna Bofa, our Chief Commercial Officer. Later on in the call, Marius Haas, our Chairman of the Board, and Jacobo Ortiz, our Chief Operating Officer, will be with us. Remarks we make today contain forward-looking statements. Such statements carry risks and uncertainties that might lead actual results to differ materially. Kindly refer to the risk factors described in our Form 10-K, Forms 10-Q, and other SEC filings. We will also talk about non-GAAP measures. You can find definitions and related disclosures in our investor presentation. Once our prepared remarks are done, we will open the line for your questions. At this point, I'll hand the call over to Lee.

Thanks, Rodrigo, and good morning to all of you. This marks my first earnings call in the CEO role. I'd like to use this opportunity to explain where Fermi is headed and the path we'll follow to get there. Let me begin by giving you some context about the conditions we're facing. Over the last few months, the market has shifted considerably. Electricity has become more constrained. Access to electrons has been narrowed by tighter moratoriums across individual states, and take Texas and ERCOT as an example, where regulators halted approvals for new large load hookups, which are what supply AI data centers with power. Finding land has also grown more difficult. Many potential sites have been ruled out because of community pushback over data center expansion, worries regarding water, noise, light, and the natural environment. Obtaining permits has become tougher as well. Take Texas: before the election, it stopped issuing new air permits, and last week's elections introduced even more uncertainty. So clearing those gates to first power became tougher. That is pushing up demand for safe harbor sites, where the area gains jobs, economic development, and long-term investment. As a result, investment-grade hyperscalers are coming to us. We have also observed changes in the way AI infrastructure is financed. The high-yield bond market has gotten tighter. In our view, more hyperscalers are shifting away from neoclouds and toward direct relationships. While that shift happens, we are concentrating our commercial efforts on deals made directly with investment-grade hyperscalers. Given that background, we issued two announcements this morning. The TensorWave lease will not proceed. That transaction was a complicated Turnkey lease involving a neocloud, several parties, and third-party credit support. The big news today: we've signed an exciting agreement with an investment-grade hyperscaler covering 600 MW of our 2027 power capacity. This one is simpler — power goes to a single customer, and there's less to finance. By every measure that matters to us, it's a stronger deal. Capacity is nearly triple. The counterparty is an investment-grade hyperscaler dealing with us directly, so no one else's credit is in play. Because the customer is constructing its own data centers at our Project Matador, Fermi doesn't bear that construction capital requirement. And notably, it's providing capital up front to speed construction. The package includes a prepayment agreement worth $200 million, which gives us significant funding up front for the 600 MW development, along with an early construction and site development agreement that lays out a framework for beginning site preparation. We're enthusiastic about how this deal is structured, and it involves little or no dilution. That's our strategy in action. By investing early to get the site ready, we help customers build more quickly, draw their capital in sooner, and convert their power and infrastructure advantages into long-term revenue opportunities. We're still working toward a definitive long-term lease, and the customer's identity remains confidential. What sets us apart is not merely having power on hand — it's the effort we've invested so customers can convert that power access into data centers running sooner. That effort carries through to the property beneath us. We're building our campus on acreage leased from the Texas Tech University System. Given conditions in this market, that relationship matters more than ever. This week, Texas Tech Chancellor Brandon Creighton restated that the university, quote, "remains firmly committed to the success of Project Matador," unquote. At a time when local pushback is eliminating candidate sites, that sort of institutional support is, in our view, a key reason our location serves as a secure harbor. We hold the power, the land, and the permits while many rivals run into strong headwinds. Over $1.6 billion of our money has gone into the site; we've completed the long lead work before lease construction, engineering design, site preparation, and moving equipment to site, which lowers execution risk for our customers and our lenders; on permits, air permits for 6 GW are secured. A comprehensive water strategy built around conservation is in place. We designed the project to minimize water use by maximizing dry and hybrid closed-loop system. At full capacity, our water usage should be roughly 80% lower compared with traditional methods. On the fuel side, Energy Transfer has committed as much as 220,000 MMBtu of natural gas daily. We've put in 4.6 miles of onsite pipeline with a rating of 450 million cubic feet per day — ample for 2 GW. As for power, we supply it ourselves. Behind the meter, on a grid we own, we've secured generation equipment exceeding 2 GW. A good number of people in the Panhandle community back this project. And finally, while rivals find it hard to obtain equipment, our turbines are already onsite; such advantages are difficult for others to match. We have a solid customer pipeline that recognizes this opportunity. Thanks to our careful capital management, we're well positioned to move Fermi beyond its pre-revenue, pre-lease stage into the next phase. At this point, I'll hand things over to Anna, who will explain how we're developing our customer portfolio.

Thanks, Lee. I'm genuinely excited about the commercial momentum we're building at Fermi, and I'd like to take a few minutes to explain why this hyperscaler relationship matters so much and to share what we're seeing across the wider customer pipeline. Power has emerged as one of the key constraints on AI growth, and the question is no longer simply how much of it you can deliver. What matters is speed, reliability, and whether the infrastructure exists to support development at scale. One point keeps coming through in our discussions with the largest technology companies in the world. Rather than waiting for power to reach them, customers now build around locations where power is already available. The 600 MW commitment Lee just walked through illustrates that shift well; we've worked closely for months with this customer's commercial and technical teams, hearing their requirements, learning their development timelines, and shaping our plans around what bringing capacity online as fast as possible will require. What's especially exciting is how we've managed to align around each party's strengths. These are sophisticated customers. They bring their own engineering standards, building designs, construction teams, and capital. Building their data centers for them isn't necessarily what they need from us. Reliable power, development-ready land, and a partner that helps them move faster are what they need. That's exactly the purpose behind our Powered Land+ offering. Fermi supplies the power, the land, and the supporting campus infrastructure. Customers can bring their own designs and construction capabilities, which lets us move forward together faster without Fermi having to finance and deliver the whole data center. Here we've gone a step further. Engineering and design work with our partners is already advanced, and we locked in critical long-lead materials before the customer's final lease. Structural steel is a good illustration. Instead of starting a fresh procurement cycle, this customer can draw on steel we've already secured to speed up its construction schedule. That's the value of the infrastructure work completed in advance, and it's why being ready counts for so much in this market. The preparations ahead of finalizing the long-term lease. To me, that reflects both how urgent demand is and how strong the partnership we've built is; we're targeting initial power availability in late Q2 or early Q3 of 2027, and the customer is aligning its development plans with that schedule. For Fermi, this opens the opportunity to bring our power capacity into revenue sooner while deploying less capital into customer-owned data center construction. Importantly, we believe this approach can be replicated. Our pipeline is strong, with active discussions across multiple investment-grade hyperscalers. Should the opportunities we're pursuing today convert into definitive agreements, they could absorb our planned 2.2 GW of capacity, and our broader pipeline extends several gigawatts beyond that. Beyond that, we’re seeing customers consider much more than their first deployment; they’re asking about rights to expand later and how they can grow alongside us as more power comes online. That sits at the heart of what we have in mind for Project Matador. Our approach is to secure the campus around 4 to 5 investment-grade hyperscalers, with each one able to set up an initial presence and then expand as time goes on. That opens the possibility of a diversified long-term customer portfolio, stronger visibility into demand, and a more disciplined way of planning and financing future infrastructure. Right now our emphasis is execution: turning our pipeline into financial commitments and long-term agreements, matching customer construction schedules to our 2027 and 2028 power ramp, and getting capacity into revenue as fast and as responsibly as we can. At the end of the day, we aren’t merely selling megawatts. We’re building the infrastructure and partnerships that let some of the world’s largest technology companies move faster toward their ambitions. We have laid the foundation. Now customers are starting to build on it. With that, I’ll hand it to Rob to explain how our three commercial offerings and wider capital strategy support that growth.

Thanks, Anna, and good morning. I'll go over our results for the third quarter, along with our liquidity, our capital strategy, and our covenants today. Before getting into that, though, I'd like to expand on what Anna said and discuss the various ways Fermi can generate differentiated value. During this quarter, we laid out our business strategy, which brings our pricing, cost, and capital requirements into alignment over the long term. That strategy is visible in action through the recent news, which lines up closely with the current market, the environment for raising capital, and what our customers require. For this early stage of our business, it's the right approach. Under one triple net lease, Fermi creates value in two ways: we sell private power behind the meter, and on our campus we lease data center real estate. Each of those channels has a fixed charge and a variable charge. The fixed charges — the capacity charge for power and the base rent for real estate — let us recover capital and costs and earn our return. Under our take-or-pay arrangements, both escalate over terms of 15 to 20 years. The variable charges for energy and common area maintenance pass fuel, operating costs, insurance, and taxes through to customers. Power operates identically in every deal. What varies is how much real estate we build. Fermi offers customers three options. We're concentrating on Powered Land+, because that's the offering the hyperscalers we're in discussions with keep asking for. Under the Powered Land+ model, what the customer gets is a parcel that's ready for development, with power, water, fiber, security, and shared campus infrastructure already in place. The data center itself is built and operated by the customer. Hyperscalers bring their own design and build partners. Of all the options, Powered Land+ can be delivered fastest. Given that we're a private power company, this model suits us well. What sets us apart from others in a crowded data center market is dedicated power on our campus. With Powered Land+, Fermi puts in the least amount of capital; that approach channels money toward the areas where we hold an edge—power, fuel, and permits—while the customer keeps the construction risk, and that sharply reduces how much capital we need. To give you a way to think about each model, page 11 in our investor presentation lays out economics taken from recent market transactions. These serve as market benchmarks, not as Fermi deal terms; they are based on market rates drawn from 45 deals over a two-year span. They might not mirror where the market sits right now, and just a small share of them are behind-the-meter deals. The point that counts is relative value, even at these historical rates. Based on the data, building Powered Land+ runs somewhere between $3.9 million and $5.1 million for each MW of total power capacity. At rates the market currently offers, that translated into fixed charges of around $55 to $69 per kW month, covering base rent plus the capacity charge. Then there's our second model, Powered Shell, which includes all of Powered Land+ — we additionally provide the building along with its core infrastructure, while the customer handles fit-out, leaving capital and revenue somewhere in the middle. Recent deals show Powered Shell costing approximately $5.1 million to $6.6 million per MW of total power capacity to construct. At historic market rates, it brought in fixed charges of roughly $64 to $80 per kW month, again including base rent and capacity charge. Powered Shell brings in more revenue for each megawatt than Powered Land+ does, but it calls for a larger amount of Fermi capital, and we take on the extra construction and delivery risk that comes with the building. The third model is Turnkey; with Turnkey we handle development, construction, and delivery of the full facility. It produces the highest contracted revenue per megawatt, though in recent market deals Turnkey has needed roughly 3x the capital that Powered Land+ requires for approximately 2.4x the revenue. Since each deal is negotiated according to what the customer needs, the terms will differ from one deal to the next. When it comes to capacity for 2027 and 2028, Fermi and the customers we may serve are concentrating on acting fast, putting money in early, and constructing their own sites; that is Powered Land+, and it holds down what we need to invest. That arrangement enables us to sell electricity once it becomes available without bearing the full construction funding on our own books. The people we serve receive the electricity they need within the schedule they want, and Fermi is able to earn revenue in step with power delivery. Now, a brief look at the results for our third quarter. This morning my focus will be cash and liquidity, because the quarter's details are in the release; we closed the quarter with roughly $316 million in cash equivalents plus restricted cash, which includes $286 million of unrestricted cash. Property, plant and equipment rose by approximately $120 million during the quarter, reaching about $1.67 billion. At quarter end, and after our July offering of $430 million in convertible senior notes, our debt stood at $958 million, net of discounts and issuance costs. Regarding liquidity, we are exercising discipline in capital deployment, aligning outflows more closely with inflows. In Q3, operations used $54.5 million in cash, while capital expenditures came to $133 million — a decrease from $185 million in Q2, with equipment facilities providing partial support. Looking ahead, we carry $126 million in equipment commitments; $112 million of that total will be covered through our facilities across the rest of 2026, through 2027, and on into 2028. As we shift toward vertical construction to meet lease requirements, the funding approach is shifting too — moving away from Fermi corporate cash and toward customer prepayments along with project-level debt and equity — which means our budgets and spending track the project resources, reducing the strain on corporate cash burn. Turning now to our capital strategy. Throughout the campus, our approach to capital is designed to lower the amount of Fermi equity required per megawatt and to cut the corporate capital we have at risk as capacity grows, and we'll accomplish this in three ways. First, employ debt and equity at the project level while keeping recourse to our corporate balance sheet limited. Second, give priority to Powered Land+ transactions for the 2027 capacity and the development planned for 2028. Third, draw customer capital in sooner via development contributions, reservation fees, prepayments, and mobilization payments, such as those we mentioned earlier. We'll provide updates on financing and capital commitments at each milestone we reach. A quick word on covenants. We've made solid progress during the last couple of months. To begin, the turbine warehouse facility with MUFG. During October, we helped move that facility over to a fresh group of lenders. This amendment brings Fermi four meaningful advantages; it did away with the November 10th lease deadline. It also removed any risk that our lenders would market the turbines for sale once November 10th passed. And it held our minimum principal payment at 5% of the loan balance instead of stepping up to 10%. According to our estimates, that avoids potential required principal payments exceeding $55 million across the next three quarters, cash that stays in the business. Finally, it allows us to relocate the equipment to the site, which cuts extra storage, transport, and insurance costs from keeping it elsewhere. Because that relocation is essentially complete, we're now able to start building the turbines vertically without delay; as part of that arrangement, we paid the lenders approximately $25 million in amendment fees. That said, I'll turn things back over to Lee.

The position we’re in is an advantaged one because of what we’ve built. We have permits, power equipment on site, fuel, water, and the site work, plus an impressive list of proven contractors. Demand for what we offer is strong. Near term, we have five priorities. First, finalize a definitive lease agreement with our hyperscaler that we talked about this morning. Second, convert our other customers in our pipeline into tenants. Third, protect liquidity and match spending to commitments. Fourth, advance financing partnerships and bring in customer capital. Fifth, keep the site ready and work safely so we can deliver first power next year. I want to thank our employees and contractors. They’re the people really installing the turbines, laying pipe in the ground, and securing our permits. I want to express gratitude to Texas Tech University and the Amarillo community. I also want to thank our investors. We are fully dedicated—100%—to turning our strengths into cash flow and returns for you. Rob, Anna, and I have now been joined by our Chairman, Marius Haas, and our Chief Operating Officer, Jacobo Ortiz, and we’ll take whatever questions you have.

Thank you so much for the question, and congratulations on the new hyperscale relationship. I'd like to ask about the sort of pending items when it comes to executing the definitive lease here. What are the main negotiating points, and how are you thinking about folding in perhaps other elements of your forward purchasing of long lead time items as you build out that relationship further beyond this initial 600 MW? Thank you.

Hi, this is Anna. Glad to take that question. Regarding where we stand with the lease negotiations, we are very, very far along. As the announcements this morning indicate, we are actively pursuing the schedule, and that means we have to begin mobilizing on our site so we can hit the timeline for our power and so the customers can hit their goals around getting their data centers stood up. What that tells you is we are getting very far along, and both sides are committed to working through the final items for the lease. As soon as that is done, we will of course announce it. We are also actively working on not just the 600 MW opportunity but additional opportunities for capacity in 2028. The goal is not just to do the initial phase but to have a second phase following that. Again, we will announce those definitive agreements as soon as they are signed and we are able to give you more details.

Thanks for that, Anna. Just as a quick follow-up, I was hoping to get a bit more detail on any permits the hyperscaler might require for the structures they'd be looking to build. I understand you hold the TCEQ air quality permit for up to the 6 GW, but I'm wondering whether any additional permitting would still be pending under this arrangement. Thank you.

Yes, we've worked very closely with them on every permitting requirement. One advantage on our side is that the heavy lifting on permitting has already been done, so our customers can take advantage of that and lean on the work we've completed. In fact, that's been one of the things that has really sped up the conversations, because we've had that permitting just as we already have this deal. It simply lets us move quickly and in lockstep with the customer.

Great. Thanks.

Hey, good morning, everyone. I wanted to ask about that $200 million prepayment as well. Rob, you mentioned it, and Anna brought it up too, in connection with getting construction started vertically on the site. So is it fair to say the Texas Tech 99-year lease is now locked in? Because we had been under the subject NTP with the firm lease agreement. Sorry to interrupt.

This is Anna again. One thing Lee emphasized earlier was how strong our relationship with Texas Tech is. They've been working closely alongside us so that every covenant tied to that lease gets satisfied, and as partners in that effort they've been outstanding. In the same way, the hyperscale customer has partnered wonderfully with both us and TTU. All three parties have collaborated to make sure TTU's requirements and that covenant are fulfilled, and we have every expectation of achieving that.

No, that's clear. I suppose, just returning to, I suppose, that $200 million prepayment, I suppose what does that actually, I suppose, let you speed up, when and where, and just kind of how fast, I suppose, how fast does that actually move the process along?

Yeah, so it lets us speed things up enormously, and what we've been doing alongside the customer is really working out when their data centers can be delivered and when our power can be ready to line up with that. So together we settled on what amount of capital we'd require before securing that definitive lease and completing the project financing, to make sure we could meet that schedule jointly. So that's where that roughly $200 million figure came from, from us collaborating to settle on that amount. What it enables us to do is we aim to begin mobilizing on site by December 1st, and so with this payment received, we'll be in a position to move forward and get started on that.

Perfect. Thanks, Anna.

Thanks.

Great, thank you. A handful of quick follow-ups, and I'll aim these at Anna. First, Anna, on the Texas situation and the permits involved: can that prospective tenant break ground immediately, or are we still waiting on something else on the permitting side before they're able to start building? Second, going by your team's outlook, when would you expect the tenant to begin taking power off the site, and is there any additional detail you can share regarding — you mentioned the 200 MW of grid power — could you walk us through your plans for getting the on-site assets up and running? You put a lot of detail on your slides, but could you give any color on what you'd expect to hand over to the tenant at those dates.

Absolutely. This is Anna again, and I'll cover the permitting portion of your question first. To be clear, what we have at our site is every permit obtained so far that a customer needs in order to begin construction and take power. We've therefore been able, working alongside the customer, to make sure that all of their requirements — from environmental to air permitting to local permitting — are satisfied, and we've accomplished that jointly. On the second question about offtake and timing: as I said, our current plan lays out a target of late Q2 to early Q3 for their first building to come online, and after that we'll keep delivering buildings — as many as eight — as they're ready. Our power ramp will be in place for them. The aim is for this to happen in 2027.

Understood. Thanks.

Hello, and thanks for taking my questions. First, regarding the new deal, I'm curious what you can share about possible lease terms and renewals, and what the duration of a lease like this would be. Second, perhaps pulling back a bit from this deal and considering the worth of the underlying assets you already own, specifically the gas turbines. There have been a number of headlines about the value of turbines and engines, like the Anthropic news lately, and I'm wondering whether you could or would be willing to comment on whether you've had any approaches for your generation assets, and what you think the Fermi year-end fleet might fetch on the secondary market right now. Thanks.